Planning Climatic Tests Using Physics-of-Failure Principles and Empirical Models

When one faces the need to plan a climatic test (temperature, temperature-humidity, rain, solar radiation etc.), several options are available. The most frequently implemented option is the use of standards. Often, the use of standards leads to testing programs that do not reflect the system's life cycle, or that prevent differentiating between systems with different life cycles. This is true even for standards that implement the tailoring philosophy. How can climatic tests be tailored to reflect true life time such as 10,000 hours or 100,000 hours? A methodology that makes use of physics-of-failure principles and empirical models provides a more realistic solution to this problem. Using empirical models to describe the environmental loads, and damage accumulation models under different loads, the effects of real life can be simulated and compared with the effects of the simulated testing conditions. The paper describes the use of this methodology. Empirical models are used to describe diurnal thermal and humidity cycles. An empirical model of temperature distribution is applied to determine the duration of exposure to varying temperature during a given life cycle. Several damage accumulation models under cyclic thermal loads are compared. One is a general power model with changing exponents for different materials. An additional model relates to the behavior of solder joints. The application of the models is compared for two different conditions, one for which the different daily temperature changes are considered, and one for which monthly average temperature changes are used to describe the daily conditions for a certain month. To evaluate temperature-humidity tests, models that describe corrosion damage under temperature-humidity and humidity penetration models are implemented to evaluate the effects of the testing conditions, relative to the real life. The advantages and problems in the implementation of the methodology are discussed in the summary of the paper.